There have been proposed various techniques to properly perform the load distribution in a system where information processings are distributed to plural processing mechanisms. In the distribution processes, not only the information processing may be assigned but also the assignment may be shifted (transited) depending on the operational state of the system.
As an example of the load distribution technique, in a system where plural information processing sections perform information processings which are distributed, a technique is employed where an agent selects the information processing section having a low operating rate and assigns a processing request to the selected information processing section. In this technique, the information processing section transmits a reassignment request for reassigning the processing request to the agent based on the processing state of the information processing section. Upon receipt of the reassignment request, the agent transfers a processing request managed by another information processing section to the information processing section transmitted the reassignment request.
As another example, in an object-oriented client-server system, the load states of the servers are managed, and based on the load states, a server object is transferred to the server having less load. Further, there has been proposed another technique in which, in distributed platforms, when an application is executed, the application and the like are distributed based on a predetermined algorithm.
Further, there has been proposed another technique in which when a request from a client is assigned (transferred) to any of a plurality of data centers on a network, the request is preferentially assigned to a data center having less processing delay time determined based on the position of the data centers and the communication routes to the data centers.
FIG. 1 illustrates a specific example of a load distribution system including a plurality of transfer processing devices transferring data to a predetermined transfer destination device, the data being transmitted from the transfer source device of the data. In FIG. 1, in transfer source devices 802 which are sources to transfer data, corresponding Virtual Machines (VM) are configured. The transfer source device 802 extracts VM statistical information relative to one or more VMs every predetermined cycle, the VMs operating in the transfer source device 802. Further, the transfer source device 802 transmits the extracted VM statistical information to a distributing device 806. The distributing device 806 serves as a so-called load balancer. Namely, the distributing device 806 transmits the received VM statistical information to one transfer processing device 804 that assigns the process of the VM statistical information of the transfer source device 802 from among a plurality of the transfer processing devices 804. The transfer processing devices 804 separates the VM statistical information for each of the VMs, and distributes and transmits the VMs to the respective transfer processing devices 804. The transfer processing devices 804 perform statistical processing for each of users who are the supply targets of the VMs. Further, a transfer management device 805 manages an assignment state in a data transfer process of the transfer source device 802 corresponding to the transfer processing devices 804.
In this example of the load distribution system, depending on the processing states of the transfer processing devices 804, the assignments of the data transfer process of the transfer source device 802 is shifted from one transfer processing device 804 to another transfer processing device 804.
For related art, reference may be made to Japanese Laid-open Patent Application No. 09-231184, Japanese Laid-open Patent Application No. 2000-10936, and Japanese Laid-open Patent Application No. 2004-520641.